Research on the construction of the prediction model of economic management students' performance under machine learning technology

In order to explore the predictive model of economic management students' performance based on machine learning techniques, three algorithms, namely, Random Forest, Support Vector Regression (SVR) and Long Short-Term Memory Network (LSTM), are used for modeling. The dataset was obtained from 2,000 economics and management students from eight universities across China and contained 28-dimensional features, such as class attendance rate (92.5% ± 5.8%), assignment submission rate (94.2% ± 4.1%) and MOOC access frequency (2.7 ± 1.3 times per day). Through data preprocessing, feature selection and dimensionality reduction, the LSTM model was finally selected as the best prediction model, with a mean square error (MSE) of 2.73, a root mean square error (RMSE) of 1.65, and a coefficient of determination, R², of 0.89. Compared with the other models, the LSTM had a higher prediction accuracy in the high performance group (MSE = 2.51), and showed better stability and adaptability. The results of the study show that machine learning techniques can effectively capture the complex relationship between student performance and learning behaviors, and provide an accurate method for predicting student performance in the field of education.